The genetic characterization of a growing number of diseases has spurred efforts to correct dysfunctions at the molecular level by gene therapy. One approach is to transfect skin cells, such as keratinocytes and fibroblasts, with genes which secrete therapeutic proteins. The ability of these secreted proteins to localize to the target will dictate the therapeutic efficiency. The long-term objective of this project is to define the molecular parameters that determine a secreted protein's ability to either remain at the site of production, or to rapidly enter the blood. A model system will be used which consists of cytokine secreting and utilizing cells. Epidermal growth factor (EGF) will be used as a model cytokine. Cells will be constructed that secrete a series of EGF molecules modified by site-directed mutagenesis. Modifications will include amino acid changes that affect receptor binding and amino terminus extensions that affect EGF size and charge. An increase in the size of EGF is predicted to significantly increase its concentration in the blood by decreasing both local tissue retention and urinary excretion. Cells secreting modified EGF molecules will be used together with cell lines expressing a series of modified EGF receptors to create artificial "paracrine" and "autocrine" systems. Cells will be implanted into artificial extracellular matrices in which all important parameters can be independently modified. The ability of modified ligands to traverse cell-containing matrices will be systematically explored using a continuous flow diffusion apparatus. This information will be used to construct realistic computer models of the ability of cytokines and other secreted proteins to translocate from the site of production into the circulation. Cells secreting modified EGF molecules will be transplanted into animals using a skin-equivalent system. The ability of the different molecules to enter the circulation in-vivo will be compared with the in-vitro model predictions, thus allowing for the development of a more realistic in-vivo model. The biological effectiveness of the different modified cytokines (and hence the effectiveness of the delivery strategy) will be evaluated by their ability to support the growth of EGF-dependent tumors in vivo at sites both proximal and distal to the skin equivalent. These studies will provide an essential foundation for identifying effective strategies for gene therapy using transkaryotic implantation. In addition they could lead to realistic animal models for diseases in which paracrine and autocrine signaling in the skin is aberrant.